Analogue sandbox models are important in the study of reservoir geology because they can offer insight into geological processes that we are rarely able to observe in nature. Seismic physical modelling is used to study the effects of seismic wave propagation in isotropic and anisotropic media and is particularly suited to isolating the effects of a single parameter independently from all others in an infinitely complex geological system. Seismic physical modelling has also been used for the testing of numerical processing algorithms, aid to evaluate interpretations of field seismic sections with scaled representations of geological formations. For this project, I set about developing methods to combine these two independent modelling techniques for the first time. However, previous attempts to use sand as a seismic modelling material failed due mainly to problems with understanding and controlling the distribution of the grain packing.This research has addressed a number of these problems through systematic laboratory experimentation that has provided new insight into the factors that affect unconsolidated sediment acoustics. An innovative technique of recording seismic physical modelling surveys has been developed so that it is now possible to successfully record ultrasonic reflections within analogue sandbox models in three-dimensions (3-D), providing benefits for both analogue sandbox and seismic modelling disciplines. For sandbox modelling, the recording of seismic images allows more detailed analyses of the structures than previously possible. For seismic modelling, more geologically realistic settings can be modelled at a fraction of the cost and construction time of conventional models. However, the greatest benefit of this new technology is that it is now possible to build seismic physical models from porous media, rather than solid, non-porous materials that ++ / are conventionally used. This scientific advance allows different fluids to be incorporated into physical models for the first time.Time-lapse 3-D seismic is becoming increasingly important in the management of hydrocarbon production, yet there is a lack of model data to support some of the conclusions being deduced. The controlled physical modelling laboratory environment combined with the ability to consistently repeat the 3-D seismic survey process now allows time-lapse seismic experiments to be performed without the need for the costly and time consuming data processing that is necessary to match legacy 3-D field data. This subsequently avoids any pitfalls that may be associated with the process, such as the masking of true fluid flow anomalies or the generation of false anomalies from data acquisition footprints.A series of time-lapse models are presented where the three-dimensional movement of fluids through the models is remotely monitored using time-lapse 3-D seismic data. These models demonstrate the true seismic response that comes from recording real data from models that undergo real changes representative of reservoir environments. Such models are inexpensive and allow rapid data turn around in a matter of days. The techniques developed here provide a new research tool that can be used to improve our understanding of the dynamics of fluid flow within porous sediments, or to study the seismic response of reservoirs as they change with time.

Natural gas hydrate, a type of inclusion compound or clathrate, are composed of gas molecules trapped within a cage of water molecules. The presence of gas hydrate has been confirmed by core samples recovered from boreholes. Interests in the distribution of natural gas hydrate stem from its potential as a future energy source, geohazard to drilling activities and their possible impact on climate change. However the current geophysical investigations of gas hydrate reservoirs are still too limited to fully resolve the location and the total amount of gas hydrate due to its complex nature of distribution. The goal of this thesis is twofold, i.e., to model (1) the heterogeneous gas hydrate reservoirs and (2) seismic wave propagation in the presence of heterogeneities in order to address the fundamental questions: where are the location and occurrence of gas hydrate and how much is stored in the sediments. Seismic scattering studies predict that certain heterogeneity scales and velocity contrasts will generate strong scattering and wave mode conversion. Vertical Seismic Profile (VSP) techniques can be used to calibrate seismic characterization of gas hydrate expressions on surface seismograms. To further explore the potential of VSP in detecting the heterogeneities, a wave equation based approach for P- and S-wave separation is developed. Tests on synthetic data as well as applications to field data suggest alternative acquisition geometries for VSP to enable wave mode separation. A new reservoir modeling technique based on random medium theory is developed to construct heterogeneous multi-variable models that mimic heterogeneities of hydrate-bearing sediments at the level of detail provided by borehole logging data. Using this new technique, I modeled the density, and P- and S-wave velocities in combination with a modified Biot-Gassmann theory and provided a first order estimate of the in situ volume of gas hydrate near the Mallik 5L-38 borehole. Our results suggest a range of 528 to 768×10^6 m^3/km^2 of natural gas trapped within hydrate, nearly an order of magnitude lower than earlier estimates which excluded effects of small-scale heterogeneities. Further, the petrophysical models are combined with a 3-D Finite Difference method to study seismic attenuation. Thus a framework is built to further tune the models of gas hydrate reservoirs with constraints from well logs other disciplinary data.

Natural gas hydrate, a type of inclusion compound or clathrate, are composed of gas molecules trapped within a cage of water molecules. The presence of gas hydrate has been confirmed by core samples recovered from boreholes. Interests in the distribution of natural gas hydrate stem from its potential as a future energy source, geohazard to drilling activities and their possible impact on climate change. However the current geophysical investigations of gas hydrate reservoirs are still too limited to fully resolve the location and the total amount of gas hydrate due to its complex nature of distribution. The goal of this thesis is twofold, i.e., to model (1) the heterogeneous gas hydrate reservoirs and (2) seismic wave propagation in the presence of heterogeneities in order to address the fundamental questions: where are the location and occurrence of gas hydrate and how much is stored in the sediments. Seismic scattering studies predict that certain heterogeneity scales and velocity contrasts will generate strong scattering and wave mode conversion. Vertical Seismic Profile (VSP) techniques can be used to calibrate seismic characterization of gas hydrate expressions on surface seismograms. To further explore the potential of VSP in detecting the heterogeneities, a wave equation based approach for P- and S-wave separation is developed. Tests on synthetic data as well as applications to field data suggest alternative acquisition geometries for VSP to enable wave mode separation. A new reservoir modeling technique based on random medium theory is developed to construct heterogeneous multi-variable models that mimic heterogeneities of hydrate-bearing sediments at the level of detail provided by borehole logging data. Using this new technique, I modeled the density, and P- and S-wave velocities in combination with a modified Biot-Gassmann theory and provided a first order estimate of the in situ volume of gas hydrate near the Mallik 5L-38 borehole. Our results suggest a range of 528 to 768×10^6 m^3/km^2 of natural gas trapped within hydrate, nearly an order of magnitude lower than earlier estimates which excluded effects of small-scale heterogeneities. Further, the petrophysical models are combined with a 3-D Finite Difference method to study seismic attenuation. Thus a framework is built to further tune the models of gas hydrate reservoirs with constraints from well logs other disciplinary data.

Kyoto University (京都大学) / 0048 / 新制・課程博士 / 博士(工学) / 甲第14138号 / 工博第2972号 / 新制||工||1441(附属図書館) / 26444 / UT51-2008-N455 / 京都大学大学院工学研究科社会基盤工学専攻 / (主査)教授 松岡 俊文, 教授 石田 毅, 教授 三ケ田 均 / 学位規則第4条第1項該当

Seismic imaging is a technique that uses seismic echoes to map and detect underground geological structures. The conventional seismic image has the resolution limit of λ/2, where λ is the wavelength associated with the seismic waves propagating in the subsurface. To exceed this resolution limit, this thesis develops a new imaging method using resonant multiples, which produces superresolution images with twice or even more the spatial resolution compared to the conventional primary reflection image. A resonant multiple is defined as a seismic reflection that revisits the same subsurface location along coincident reflection raypath. This reverberated raypath is the reason for superresolution imaging because it increases the differences in reflection times associated with subtle changes in the spatial location of the reflector. For the practical implementation of superresolution imaging, I develop a post-stack migration technique that first enhances the signal-to-noise ratios (SNRs) of resonant multiples by a moveout-correction stacking method, and then migrates the post-stacked resonant multiples with the associated Kirchhoff or wave-equation migration formula. I show with synthetic and field data examples that the first-order resonant multiple image has about twice the spatial resolution compared to the primary reflection image. Besides resolution, the correct estimate of the subsurface velocity is crucial for determining the correct depth of reflectors. Towards this goal, wave-equation migration velocity analysis (WEMVA) is an image-domain method which inverts for the velocity model that maximizes the similarity of common image gathers (CIGs). Conventional WEMVA based on subsurface-offset, angle domain or time-lag CIGs requires significant computational and memory resources because it computes higher dimensional migration images in the extended image domain. To mitigate this problem, I present a new WEMVA method using plane-wave CIGs. Plane-wave CIGs reduce the computational cost and memory storage because they are directly calculated from prestack plane-wave migration, and the number of plane waves is often much smaller than the number of shots. In the case of an inaccurate migration velocity, the moveout of plane-wave CIGs is automatically picked by a semblance analysis method, which is then linked to the migration velocity update by a connective function. Numerical tests on synthetic and field datasets validate the efficiency and effectiveness of this method.

This thesis considers the reconstruction of subsurface models from seismic observations, a well-known high-dimensional and ill-posed problem. As a first regularization to such a problem, a reduction of the parameters' space is considered following a truncated Discrete Cosine Transform (DCT). This helps regularizing the seismic inverse problem and alleviates its computational complexity. A second regularization based on Laplace priors as a way of accounting for sparsity in the model is further proposed to enhance the reconstruction quality. More specifically, two Laplace-based penalizations are applied: one for the DCT coefficients and another one for the spatial variations of the subsurface model, which leads to an enhanced representation of cross-correlations of the DCT coefficients. The Laplace priors are represented by hierarchical forms that are suitable for deriving efficient inversion schemes. The corresponding inverse problem, which is formulated within a Bayesian framework, lies in computing the joint posteriors of the target model parameters and the hyperparameters of the introduced priors. Such a joint posterior is indeed approximated using the Variational Bayesian (VB) approach with a separable form of marginals under the minimization of Kullback-Leibler divergence criterion. The VB approach can provide an efficient means of obtaining not only point estimates but also closed forms of the posterior probability distributions of the quantities of interest, in contrast with the classical deterministic optimization methods. The case in which the observations are contaminated with outliers is further considered. For that case, a robust inversion scheme is proposed based on a Student-t prior for the observation noise. The proposed approaches are applied to successfully reconstruct the subsurface acoustic impedance model of the Volve oilfield.

This thesis focuses on the concept of Single Well Imaging (SWI) in which a seismic source and receivers are deployed in a borehole to investigate the surrounding geology. The Uniwell project (1997-1999) was the first attempt to develop the SWI method; it used a fluid-coupled downhole source, which unfortunately generated high amplitude guided waves in the borehole which obscured all other useful information. Initial research work detailed in this thesis focused on removing the high amplitude guided waves, known as tube waves. Two-step source signature deconvolution using first the recorded source signature, and then the tube-wave reflected from the bottom of the well, succeeded in compressing the tube wave. The results were not consistent across all receivers, but there is enough correlation to identify a P-wave. Further work concentrates on using a new technique called Empirical Mode Decomposition to separate the tube-wave mode from the data. This identifies three dominant modes and a possible body wave arrival, but the results are ambiguous due to the inability of the decomposition to focus on the narrow bandwidth of interest. The source signature deconvolution technique can also be used to process real-time vertical seismic profiling (VSP) data down-hole, during pauses in drilling, in what is referred to as a Seismic-While-Drilling (SWD) setup. Results show that the technique is versatile and robust, giving 1 ms precision on first-break picking even in very noisy data. I also apply the technique to normal VSP data to improve both the resolution and the signal-to-noise ratio. A major effort in this thesis is to consider the effect of a clamped downhole source to overcome the tube-wave problem, using a magnetostrictive source. Earlier work established that the use of a reaction mass tended to excite resonances in the tool which caused the transducer to break. A new design for the source was developed in cooperation with colleagues which utilises a hydraulic amplifier design and a low power coded waveform driving method exploiting the time-bandwidth product to extract the signal from the noise. My results show that as the run time increases the resolution improves. With a run length of 80s it is possible to resolve the signal transmitted 50 cm through a granite formation. This analysis led to a revised design of the source to improve its efficiency. I have used finite difference modelling, with a variable grid technique, to compare an ideal explosive source with an ideal clamped source. The fluid-coupled source emits high amplitude tube waves which virtually obscure the body wave, whereas the clamped source emits a clearly identifiable P-wave along with lower amplitude tube waves. This clearly illustrates the advantage of an ideal clamped source. To model the source more accurately the idealwavelet is replaced by the respective recorded source signatures, and the data is then processed by cross correlation with the appropriate signature. The results show that the coded waveform approaches the resolution of the ideal wavelet very well, with all major events being visible. However, the fluid-coupled source performs very poorly with only the highest amplitude tube-wave visible. This work illustrates that by replacing a fluid-coupled source by a clamped source driven by a coded waveform, and by processing the data using cross correlation or signature deconvolution, it is possible to minimise or eliminate tube-wave noise from a SWI survey. It is hoped that the results outlined here will provide the basis for a new SWI method than can be used to prolong the supply of North Sea oil.

Modern oil and gas exploration depends on a variety of geophysical prospect tools. One of them is reflection seismology that allows to obtain interwell information of sufficient resolution economically. This exploration method collects reflection seismic data on the surface of an area of prospect interest and then uses them to build seismic images of the subsurface. All imaging approaches can be divided into two groups: wave equation-based methods and integral schemes. Kirchhoff migration, which belongs to the second group, is an indispensable tool in seismic imaging due to its flexibility and relatively low computational cost. Unfortunately, the classic formulation of this method images only a part of the surface data, if so-called multipathing is present in it. That phenomenon occurs in complex geologic settings, such as subsalt areas, when seismic waves travel between a subsurface point and a surface location through more than one path. The quality of imaging with Kirchhoff migration in complex geological areas can be improved if multiple paths of ray propagation are included in the integral. Multiple arrivals can be naturally incorporated into the imaging operator if it is expressed as an integral over subsurface take-off angles. In this form, the migration operator involves escape functions that connect subsurface locations with surface seismic data values through escape traveltime and escape positions. These escape quantities are functions of phase space coordinates that are simply related to the subsurface reflection system. The angle-domain integral operator produces output scattering- and dip-angle image gathers, which represent a convenient domain for subsurface analysis. Escape functions for angle-domain imaging can be simply computed with initial-value ray tracing, a Lagrangian computational technique. However, the computational cost of such a bottom-up approach can be prohibitive in practice. The goal of this work was to construct a computationally efficient phase space imaging framework. I designed several approaches to computing escape functions directly in phase space for mapping surface seismic reflection data to the subsurface angle domain. Escape equations have been introduced previously to describe distribution of escape functions in the phase space. Initially, I employed these equations as a basis for building an Eulerian numerical scheme using finite-difference method in the 2-D case. I show its accuracy constraints and suggest a modification of the algorithm to overcome them. Next, I formulate a semi-Lagrangian approach to computing escape functions in 3-D. The second method relies on the fundamental property of continuity of these functions in the phase space. I define locally constrained escape functions and show that a global escape solution can be reconstructed from local solutions iteratively. I validate the accuracy of the proposed methods by imaging synthetic seismic data in several complex 2-D and 3-D models. I draw conclusions about efficiency by comparing the compute time of the imaging tests with the compute time of a well-optimized conventional initial-value ray tracing. / text

The neotectonic evolution of the eastern Mediterranean is intimately tied to interactions between the underthrusting/subducting slab along the southern margin of Anatolia and the overriding plate. The lateral variations in the subduction zone can be viewed as a temporal analogue of the transition between continuous subduction and subduction termination by continent-continent collision. By investigating the lateral variations along this subduction zone in the overriding plate, we can gain insight into the processes that precede continent collision. This dissertation summarizes the results of three studies that focus on different parts of the subduction margin: 1) In the west, where the development of a slab tear represents the transition between continuous and enigmatic subduction, 2) In the east, where continent-continent collision between the Arabian and Eurasian Plate is leading to the development of the third largest orogenic plateau on earth after complete slab detachment, and 3) In central Anatolia, where the subducting slab is thought to be in the processes of breaking up, which is affecting the flow of mantle material leading to volcanism and uplift along the margin. In the first study, we interpret that variations in the composition of material in the downgoing plate (i.e. a change from the subduction of oceanic material to continental material) may have led to the development of a slab tear in the eastern Aegean. This underthrusting, buoyant continental fragment is controlling overriding plate deformation, separating the highly extensional strains of western Anatolia from the much lower extensional strains of central Anatolia. Based on intermediate depth seismicity, it appears that the oceanic portion of the slab is still attached to this underthrusting continental fragment. In the second study, we interpret that the introduction of continental lithosphere into the north-dipping subduction zone at the Arabian-Eurasian margin led to the rollback and eventual detachment of the downgoing oceanic lithosphere attached to the Arabian Plate. After detachment, high rates of exhumation in the overriding plate are recorded due to the removal of the oceanic lithosphere and accompanying rebound of the Arabian continental lithosphere. In the third study, we image a transitional stage between the complete slab breakoff of the second study and the continuous subduction slab of the first study. We interpret that trench-perpendicular volcanism and ~2 km of uplift of flat-lying carbonate rocks along the southern margin of Turkey can be attributed to the rollback and ongoing segmentation of the downgoing slab as attenuated continental material is introduced into the subduction zone. Combining these three studies allows us to understand the terminal processes of a long-lived subduction zone as continental material is introduced.

Ray theory plays an important role in seismic imaging and velocity model building. Although rays are the high-frequency asymptotic solutions of the wave equation and therefore do not usually capture all details of the wave physics, they provide a convenient and effective tool for a wide range of geophysical applications. Especially, ray theory gives rise to traveltimes. Even though wave-based methods for imaging and model building had attracted significant attentions in recent years, traveltime-based methods are still indispensable and should be further developed for improved accuracy and efficiency. Moreover, there are possibilities for new ray theoretical methods that might address the difficulties faced by conventional traveltime-based approaches. My thesis consists of mainly four parts. In the first part, starting from the linearized eikonal equation, I derive and implement a set of linear operators by upwind finite differences. These operators are not only consistent with fast-marching eikonal solver that I use for traveltime computation but also computationally efficient. They are fundamental elements in the numerical implementations of my other works. Next, I investigate feasibility of using the double-square-root eikonal equation for near surface first-break traveltime tomography. Compared with traditional eikonal-based approach, where the gradient in its adjoint-state tomography neglects information along the shot dimension, my method handles all shots together. I show that the double-square-root eikonal equation can be solved efficiently by a causal discretization scheme. The associated adjoint-state tomography is then realized by linearization and upwind finite-differences. My implementation does not need adjoint state as an intermediate parameter for the gradient and therefore the overall cost for one linearization update is relatively inexpensive. Numerical examples demonstrate stable and fast convergence of the proposed method. Then, I develop a strategy for compressing traveltime tables in Kirchhoff depth migration. The method is based on differentiating the eikonal equation in the source position, which can be easily implemented along with the fast-marching method. The resulting eikonal-based traveltime source-derivative relies on solving a version of the linearized eikonal equation, which is carried out by the upwind finite-differences operator. The source-derivative enables an accurate Hermite interpolation. I also show how the method can be straightforwardly integrated in anti-aliasing and Kirchhoff redatuming. Finally, I revisit the classical problem of time-to-depth conversion. In the presence of lateral velocity variations, the conversion requires recovering geometrical spreading of the image rays. I recast the governing ill-posed problem in an optimization framework and solve it iteratively. Several upwind finite-differences linear operators are combined to implement the algorithm. The major advantage of my optimization-based time-to-depth conversion is its numerical stability. Synthetic and field data examples demonstrate practical applicability of the new approach. / text